status of high-priority GNATS bugs

status of high-priority GNATS bugs

[Joe Buck]

I did a scan through the 36 high-priority GNATS PRs.  Here's what I
found, hope it is useful:

Verified as present in 3.2.1-pre as of 22-oct-02 (on i686-pc-linux-gnu):

10 bugs: 5509, 6579, 6718, 6746, 6981, 7228, 7266, 7363, 7385, 7679
(Mark says "working on a fix" for 7385)

PR's 6579 and 7679 appear to be duplicates.  The infinite loop appears to
be in spew.c, and is possibly related to error recovery.

Reported against 3.2 (or earlier), not tested by me

7227 (sparc/solaris and mips/irix)
7822 (m68k-linux)
7856 (arm)

I recommend that the following 7 bugs be closed:

6745 (can't duplicate, evidently fixed)
7426 (can't duplicate, evidently fixed)
4650 in feedback since Mar, no testcase (confidential), no reply
4947 in feedback since Feb., no reply
5196 in feedback since Jan., no reply
5515 in feedback since Feb., no reply
5971 in feedback since Mar., no reply

We need more info on the following:

6545 in feedback state since Aug.
7090 it appears that a patch was applied for this one, but it is still
     in "feedback" state.
7623 (SCO, patch candidate in review as of Aug 17)

Reported against 3.3 branch, confirmed not present in 3.2 branch:

5866, 6919, 7639, 8080, 8186

Reported against 3.3 branch, not sure if 3.2 branch affected:

7794 (sparc/solaris): "regression from 3.2"
7796 (sparc/solaris)
7928 (sparc/solaris)
8085 (mips/irix)

Documentation bug, probably shouldn't be priority "high"
3386 Undocumented target macros

Misc: unclear whether this is priority "high"
6558 (wrongly reverted Ada patches?)

Re: status of high-priority GNATS bugs

[Janis Johnson]

I didn't see 6994, which is a regression from 2.95, on your list.  Zack has
a possible patch (http://gcc.gnu.org/ml/gcc-bugs/2002-10/msg00372.html) but
nothing seems to have happened with it.

Janis

Re: status of high-priority GNATS bugs

[Zack Weinberg]

On Tue, Oct 22, 2002 at 02:41:42PM -0700, Janis Johnson wrote:
> I didn't see 6994, which is a regression from 2.95, on your list.  Zack has
> a possible patch (http://gcc.gnu.org/ml/gcc-bugs/2002-10/msg00372.html) but
> nothing seems to have happened with it.

I'm working on a revised version even as we speak.

zw

Re: status of high-priority GNATS bugs

[Joe Buck]

> I didn't see 6994, which is a regression from 2.95, on your list.  Zack has
> a possible patch (http://gcc.gnu.org/ml/gcc-bugs/2002-10/msg00372.html) but
> nothing seems to have happened with it.

I changed this one to priority "high", and confirmed that the ICE is
still present in 3.2.1-pre in today's CVS.

Re: status of high-priority GNATS bugs

[Joseph S. Myers]

On Tue, 22 Oct 2002, Joe Buck wrote:

> Misc: unclear whether this is priority "high"
> 6558 (wrongly reverted Ada patches?)

The patches, including bug fixes, were applied before 3.1 branched and 
wrongly reverted after then.  ada/6919 is a similar problem (patches only 
ever applied to the branch, not mainline), also "high".

-- 
Joseph S. Myers
jsm28@cam.ac.uk

Re: status of high-priority GNATS bugs

[David S. Miller]

   From: Joe Buck <jbuck@synopsys.com>
   Date: Tue, 22 Oct 2002 13:53:30 -0700 (PDT)

   Reported against 3.3 branch, not sure if 3.2 branch affected:
   
   7794 (sparc/solaris): "regression from 3.2"
   7796 (sparc/solaris)

I know for a fact that these 3 are specific to the 3.3
sources and do not occur in 3.2

   7928 (sparc/solaris)

I do not know specifically about this, but knowing Kaveh Ghazi's
bug reporting habits, he would indicate if this affected 3.2
by either saying so in the report or openning a seperatate bug
against 3.2

To my eyes it seems clear that he is trying to say in this report
that this 3.3 report is a regression against 3.2, which I know
he tests heavily.

Re: status of high-priority GNATS bugs

[Joe Buck]

> On Tue, 22 Oct 2002, Joe Buck wrote:
> 
> > Misc: unclear whether this is priority "high"
> > 6558 (wrongly reverted Ada patches?)
> 
> The patches, including bug fixes, were applied before 3.1 branched and 
> wrongly reverted after then.  ada/6919 is a similar problem (patches only 
> ever applied to the branch, not mainline), also "high".

Thanks, then it should go into the category 

"Reported against the trunk, confirmed not present in 3.2 branch"

Right now I'm mainly interested in the releasability of 3.2.1.  Still
doesn't look good.

Re: status of high-priority GNATS bugs

[Kaveh R. Ghazi]

 > 8085 (mips/irix)

I believe Richard Sandiford is working on a fix for this one here:
http://gcc.gnu.org/ml/gcc-patches/2002-10/msg01302.html

This regression only appears on the trunk, it does not affect 3.2.x so
fixing it should not hold up the 3.2.1 release.

		--Kaveh
--
Kaveh R. Ghazi			ghazi@caip.rutgers.edu

Re: status of high-priority GNATS bugs

[Kaveh R. Ghazi]

 > 7794 (sparc/solaris): "regression from 3.2"
 > 7796 (sparc/solaris)
 > 7928 (sparc/solaris)

These regressions all occur on the trunk.  None of them fail on the
3.2.x branch so they should not hold up the 3.2.1 release.

Sorry if that wasn't clear, but I indicated these errors occured in
the 3.3 series in each PR's "release:" field.

		--Kaveh
--
Kaveh R. Ghazi			ghazi@caip.rutgers.edu

Re: status of high-priority GNATS bugs

[Mark Mitchell]

--On Tuesday, October 22, 2002 03:41:35 PM -0700 "David S. Miller" 
<davem@redhat.com> wrote:

>    From: Joe Buck <jbuck@synopsys.com>
>    Date: Tue, 22 Oct 2002 13:53:30 -0700 (PDT)
>
>    Reported against 3.3 branch, not sure if 3.2 branch affected:
>
>    7794 (sparc/solaris): "regression from 3.2"
>    7796 (sparc/solaris)
>
> I know for a fact that these 3 are specific to the 3.3
> sources and do not occur in 3.2

GNATS does not have a terribly good way of encoding this, but a
convention is arising it seems: put "[mainline regression]" in the
bug description.

Would one of you care to do this for these bugs?  That way when I
look at them, I know they aren't 3.2.1 problems, at least.

Thanks,

-- 
Mark Mitchell                mark@codesourcery.com
CodeSourcery, LLC            http://www.codesourcery.com

Re: status of high-priority GNATS bugs

[Christian Jönsson]

On Tue, Oct 22, 2002 at 05:14:06PM -0500, Brad Lucier wrote:
> I don't find 8017 in your list; it's a C++ bug that prevents the mainline
> bootstrapping at all on sparcv9-sun-solaris2.8 and has been open for almost
> a month.
> 
> I know davem said he'll look at it, but shouldn't it be on your list?

uhm, the "same" problem, according to davem, occures for
sparc64-linux. I reported on this and dave closed it, saying it's the
same as 8017.

Cheers,

/ChJ

Re: status of high-priority GNATS bugs

[Matthias Klose]

closed #7822 (m68k-linux). this one was fixed on the branch.

Re: status of high-priority GNATS bugs

[Joe Buck]

Brad writes:
> I don't find 8017 in your list;

I just changed it to "high" in GNATS.  This one does not affect the 3.2.1
branch.

Re: status of high-priority GNATS bugs

[Joe Buck]

> I don't find 8017 in your list; it's a C++ bug that prevents the mainline
> bootstrapping at all on sparcv9-sun-solaris2.8 and has been open for almost
> a month.

The list included those flagged with priority "high".  To get this
priority a bug must be a regression, but clearly failing to bootstrap on a
supported platform should count as a regression.

> I know davem said he'll look at it, but shouldn't it be on your list?

It would seem that it should be.

Re: status of high-priority GNATS bugs

[Brad Lucier]

I don't find 8017 in your list; it's a C++ bug that prevents the mainline
bootstrapping at all on sparcv9-sun-solaris2.8 and has been open for almost
a month.

I know davem said he'll look at it, but shouldn't it be on your list?

Brad

Re: status of high-priority GNATS bugs

[Richard Earnshaw]

Joe.Buck@synopsys.com said:
> To help get a sense of what remains to be done, I strolled through the
> high-priority bugs in GNATS.  Eliminating those referring to
> incomplete documentation or the Ada/ACT issues, here's what is left. 

> 5687: [arm] crash in loop.  Reported against CVS head, not sure if it
> affects 3.1 

This case is describing a problem that occurred during build/bootstrap and 
is now in state "feedback".  Also, the reported target (arm-linux) has a 
successful test run against it in the last few days.

  http://gcc.gnu.org/ml/gcc-testresults/2002-05/msg00081.html

So I don't think this is a problem any more (at least, not on the branch).

R.

status of high-priority GNATS bugs

[Joe Buck]

To help get a sense of what remains to be done, I strolled through the
high-priority bugs in GNATS.  Eliminating those referring to incomplete
documentation or the Ada/ACT issues, here's what is left.

The following seem to be in limbo, not reproduceable because of no
testcase, or waiting for feedback from people who can't send testcases:
3883, 4165, 4175, 4243, 4650, 4947

For the others, some might be gone and maybe should be closed: (?)

3373: java bootstrap issue
aoliva wrote: We should add $(FLAGS_FOR_TARGET) to GCJ_FOR_TARGET, indeed.
Is this resolved?

3589: is this resolved?  (hp illegal instr)

3954: simple returns broken when return value is a pseudo
	rth says "should be fixed in 3.1"

5196: bad makefile generated for fastjar on HP?
	(are we still seeing this, or has it been duplicated?)

As for the others:

2454: failure showing up only on v850-unknown-elf?
(if so, I guess it's not release critical)

3083: -O3/C++ too slow
(some work has been done to make this a little better, not release critical)

3931: avr-elf can't build libstdc++
	rth has proposed patch, waiting for feedback

5515: configure problem, possible shell bug on powerpc-ibm-aix4.3.3.0
	proposed patch, not sure of status

5687: [arm] crash in loop.  Reported against CVS head, not sure if it
affects 3.1

5971: make install fails for 3.0.4, java, i586 RH 7.0
	(asking for feedback)

6212: g++ EH regressions, irix6 -mabi=64
	evidently partially fixed, rth says a problem is still present

6543: ICE in extract_insn, ppc.  In feedback, proposed fix.